Head to head: Kling Image vs Ideogram V4.5 Edit
Kling Image vs Ideogram V4.5 Edit
By Ryan Merket · Published
Eight image-generation tasks put constraint-following, composition, and photographic detail under pressure—from a four-color campsite and a macro tart to a counted dumpling scene. The comparison rewards both precise prompt execution and the quality of the finished image.
Ideogram V4.5 Edit leads the saved scorecard, 62.8 to Kling Image’s 58.0, and takes seven of the eight task wins, with no ties. The edge shows up in varied prompts: Ideogram more closely follows the strict palette in the campsite, delivers the intimate macro look for the lime tart, and makes the requested ukiyo-e style more legible. The advantage isn’t flawless compliance. In the spatial-layout task, Kling followed the placements and flat-vector brief closely, while Ideogram’s room was more polished but softer and more detailed than requested. On the dumpling count, Kling won: its overhead slate-board scene better matched the brief, though it still appeared short of 17. Ideogram’s version made the count even harder to verify. These are meaningful differences across this saved task sample, not evidence of statistical confidence or a claim about every use case. The editorial result is decisive on this sample: Ideogram’s broader consistency outweighs Kling’s single task win. **Verdict: Ideogram V4.5 Edit wins this head-to-head, 7–1, on the saved task sample.**
Restricted palette
A flat-vector illustration of a mountain campsite at dusk using ONLY four colors — cream, burnt orange, teal, and charcoal — with no gradients and no stray off-palette colors, balanced composition, 16:9.
OpenAI: GPT-6 Luna: Model A has a more immersive dusk atmosphere and a lively, balanced campsite, but its multiple apparent shades and the bright turquoise clearing weaken the strict four-color requirement. Model B uses a cleaner, flatter palette and clear vector shapes, though its large empty foreground makes the composition less balanced. (Second judge pass, order swapped — scores are the average of both: Model B follows the four-color, flat-vector constraint more closely and includes a clear dusk campsite, thoug | Anthropic: Claude Fable 5.1: Model A is the more atmospheric, nicely compose…
Macro Lime Tart Focus
A macro close-up of a torched lime tart on a black stoneware plate, shot with a true macro lens and extremely shallow depth of field so only the front edge of the tart is tack sharp while the toasted meringue peaks fall into creamy bokeh, revealing microscopic zest flecks, blistered sugar, buttery crumb texture, and a thin bead of lime curd gloss; contemporary luxury dessert photography, dramatic low side light, intimate framing, 16:9.
OpenAI: GPT-6 Luna: Model B better captures the intimate macro framing, creamy shallow-focus falloff, dramatic side lighting, and glossy lime curd bead. Model A includes more of the tart and plate, but its broader, sharper presentation is less faithful to the requested macro look. (Second judge pass, order swapped — scores are the average of both: Model B better captures the intimate macro framing, dramatic low side light, black stoneware, and extremely shallow focus, with vivid crumb, curd gloss, and toasted merin | Anthropic: Claude Fable 5.1: Model B nails the requested macro look—razor-th…
Negation
A cozy reading nook with an armchair, a stack of books, and a mug of tea by a window — with absolutely NO plants, NO lamps, and NO artwork or picture frames anywhere in the frame. Warm afternoon light, 16:9.
OpenAI: GPT-6 Luna: Model A more clearly delivers the requested warm afternoon reading nook, with the chair, book stack, and tea arranged prominently by the window. Model B is also cozy and respects the exclusions, but its book stack is partly obscured and its lighting feels less warmly sunlit. (Second judge pass, order swapped — scores are the average of both: Model B more directly captures the requested nook, with a stack of books beside the armchair and a mug on the window sill; Model A is also attractive and re | Anthropic: Claude Fable 5.1: Both omit lamps, plants, and artwork indoors, b…
Named art style
A ukiyo-e woodblock print of a fishing boat riding a large cresting wave at dawn, faithful to the flat color planes, bold outlines, and stylized foam of the tradition, muted indigo and cream palette.
OpenAI: GPT-6 Luna: Model B more clearly captures the ukiyo-e woodblock tradition with crisp outlines, stylized foam, a dramatic cresting wave, and the requested muted indigo-and-cream palette. Model A has a striking wave and dawn colors, but its crowded arrangement and multiple boats make the scene less focused. (Second judge pass, order swapped — scores are the average of both: Model B more faithfully captures the requested muted indigo-and-cream ukiyo-e print, with a clearly readable fishing boat against a bold | Anthropic: Claude Fable 5.1: Model A is an almost direct derivative of Hokusa…
Count 17 Dumplings
A candid overhead food-photography scene of a midnight tasting board on dark slate, containing exactly 17 distinct crescent dumplings arranged in a loose spiral, each dumpling fully visible and individually countable, with varied pleats and tiny garnish dots of chili oil, beside a small ceramic ramekin of black vinegar, a pair of matte brass chopsticks, and scattered cilantro leaves; realistic editorial style, moody side lighting, razor-sharp focus, high contrast, 16:9.
OpenAI: GPT-6 Luna: Model A better matches the overhead slate-board scene, crescent shapes, garnish, and moody editorial lighting, though it appears to fall short of the requested 17 dumplings. Model B has a less overhead perspective, noticeably varied non-crescent shapes, and softer focus, and its dumplings are difficult to count confidently. (Second judge pass, order swapped — scores are the average of both: Model B has substantially more than 17 dumplings, with several appearing rounded rather than crescent-shap | Anthropic: Claude Fable 5.1: Model A nails nearly every element—overhead sla…
Spatial layout
A clean isometric illustration of a bedroom: a bed against the LEFT wall, a round rug centered on the floor, a desk under the WINDOW on the back wall, and a floor lamp in the FRONT-RIGHT corner. Flat-vector style, consistent perspective.
OpenAI: GPT-6 Luna: Model A follows the requested placements and flat-vector style closely, while Model B also places the bed, centered rug, desk beneath the back-wall window, and lamp in the front-right corner. Model B has a more polished, balanced composition, though its added details and softer shading are less strictly flat-vector. (Second judge pass, order swapped — scores are the average of both: Model B follows all four requested placements and presents them in a polished, consistent isometric room with espe | Anthropic: Claude Fable 5.1: Model B places the bed flush against the left w…
Crisp Jacquard Tea Setup
An ultra-detailed still life of an afternoon tea setting on a table draped with a deep indigo jacquard cloth, showing the fabric weave, frayed selvedge, embroidered pomegranate motifs, flaky croissant layers, sugar crystals, brushed silver tongs, and tiny pores on ripe apricots with uncompromising clarity; naturalistic fine-art food photography, soft window light from the left, restrained color palette, three-quarter composition, 16:9.
OpenAI: GPT-6 Luna: Model A shows the embroidered pomegranates, split flaky croissant, sugar crystals, tongs, and frayed indigo cloth especially clearly, but the arrangement feels crowded and omits a teacup. Model B has a more composed, naturalistic tea-table scene with soft window light and a visible cup, though several requested micro-details are less distinct. (Second judge pass, order swapped — scores are the average of both: Model A more clearly renders the requested jacquard weave, frayed edge, pomegranate em | Anthropic: Claude Fable 5.1: Model B delivers a cohesive, photorealistic aft…
Glass Noodle Bar Reflections
A photorealistic ramen counter scene at dusk with a clear double-walled glass bowl of amber broth and noodles in the foreground, a chrome teapot, a rain-speckled window, and a small rectangular mirror behind the condiments, all reflecting and refracting the surroundings correctly—chopsticks, neon menu glow, scallions beneath the broth surface, and distorted highlights through curved glass must read believably; cinematic urban food photography, layered composition, cool blue exterior light mixed with warm interior tungsten, 16:9.
OpenAI: GPT-6 Luna: Model A clearly shows the double-walled bowl and includes the requested counter details, but Model B creates a more convincing, cinematic rainy ramen-bar scene with stronger warm-cool lighting, steam, and reflections through the glass. Model B’s bowl is less clearly double-walled, while Model A’s chrome teapot is partially cropped, so neither matches every detail perfectly. (Second judge pass, order swapped — scores are the average of both: Model B has the more cinematic, layered food-photograph | Anthropic: Claude Fable 5.1: Model A delivers the specific technical asks—a…
Matchup powered by OpenRouter.